Inconsistent results (or something else?)

[panda4k]

Greetings!


I've been having a few MEMORY_MANAGEMENT BSODs and decided to run a memtest86, found a few errors rather quickly.

Here are the system specs to add to the context:
CPU: Ryzen 9 5900x (pbo -20 all core, vsoc 1.1v instead of 1.2v (was also tested with default pbo and vsoc))
RAM: Corsair Vengeance RGB Pro 2x16gb 3600Mhz CL18 (Hynix SK)
Motherboard: Gigabyte B550 Aorus Elite V2 (latest bios F16e)
The system was running for 9 months without issues.

I made sure to turn off xmp/docp as well and the errors remained. Since it's a 2x16gb kit, I tested both sticks individually and found that only one specific stick is having errors. I took both sticks to a friend's PC and got the same results there (exact same stick causing errors in the first 10 minutes of the test).

I've included couple pictures. The picture with only one error is with only the one "surely faulty module" that also failed in a different PC installed, running at default speeds. The picture with multiple errors is running both sticks and also on default speeds.

I've been running the one so far proven healthy stick with xmp/docp enabled for a week and then noticed certain apps not working properly and this time used HCI memtest to fill up the remaining unused memory. It found errors, I disabled xmp/docp and it resolved the errors.

Then after a week of using that one stick at default speeds (no xmp/docp) and noticed how bad performance is in certain tasks without dual channel so I put the 2nd stick back in and decided to try blacklisting bad memory in Windows as described in the documentation.

Now, it's been almost a full day of the PC running with the exact same settings as when I first noticed something related to the memory acting up (xmp/docp turned on and VSOC dropped from 1.2v to 1.1v) and it hasn't thrown any errors for multiple passes of memtest86 nor has it encountered any crashes.

I've done multiple combinations of tests with no overclocks, undervolts, different memory slots, everything stock etc. Always seemed very consistent no matter what until trying to run both sticks again when it suddenly started working without errors.

After some research, turns out a lot of people are complaining about Corsair Vengeance and their bins of memory so there is that. Also, I've reached out to the reseller already and they've told me RMA or similar will take a bit of time this time around (for some other reasons) hence I'm trying to make use of what I have for now.

Now I'm not sure anymore what's going on, perhaps it will eventually start showing errors again.
Then I noticed all errors that failed couple weeks ago are "CPU: 0" so that has me wondering if it may also be a CPU issue. But then again, the memory stick also failed at a friends PC. Does the pattern of the errors correspond to an actual faulty stick, cpu? And the obvious, why would both sticks run as advertised a week later etc, but I know this is not really possible to answer. Any input would be appreciated

[David (PassMark)]

RAM errors aren't always consistent. Can depend on temperature, EMI, the slot being used, power fluctuations, and just random chance if the stick is right on the edge of working / not working.

Sounds like you have spent a huge amount of time trying to get them to work, but I think they are faulty and need to be replaced.

If the RAM is going to take weeks, I would buy a spare 8GB stick to use in the meantime (around $20 on Newegg at the moment). Or buy another 2x16GB sticks and run 64GB when the RMA comes back.

[panda4k]

RAM errors aren't always consistent. Can depend on temperature, EMI, the slot being used, power fluctuations, and just random chance if the stick is right on the edge of working / not working.

Sounds like you have spent a huge amount of time trying to get them to work, but I think they are faulty and need to be replaced.

If the RAM is going to take weeks, I would buy a spare 8GB stick to use in the meantime (around $20 on Newegg at the moment). Or buy another 2x16GB sticks and run 64GB when the RMA comes back.

Thank you for the reply!

I did go ahead and and made sure to warm up the inside of the case and each component as much as I could, ran prime95, heavy gpu benchmarks, a lot of nvme drive usage and temperatures going up etc. I even messed around with having an old air conditioner run and have some devices on top of the case to somewhat increase EMI and electrical inconsistencies but to no avail. And during the night it passed another few memtest86 passes.

But yes, I will go ahead and continue the RMA process, thank you.
As for the replacement kit until this one is done with RMA, unfortunately I live at a place that has no access to newegg nor amazon so the local stores have their components priced quite a bit higher than needed. Perhaps I will be able to get a temporary replacement kit from the reseller, they were kind enough once already to do that when my previous nvme drive was failing.

Thanks once again!